Of the Tidy Tuesday datasets, we chose to work with a dataset that contains IMDb TV Ratings from 1990 to 2018. We have average ratings for each show’s individual seasons, as well as information about when the season first aired and genre descriptions for each show. Here, we present four plots:
1. An interactive data table showing the Top 10 Shows per Year,
2. A visualization of genre bigrams (pairs of genres for each show),
3. Violin plots showing ratings by genre, and
4. A bar plot showing whether shows that were renewed for a second season actually have better season 1 ratings.
The data table includes every show that had the highest 10 ratings by year between 1990 and 2018. One can search a year to see the top 10 shows of that year or search a show to see what year that show appeared in the top 10 as well.
Nodes denote genres, size of nodes denote total number of times geanre appeared, edge thickness indicates number of co-occurrence. The four networks are divided by decade in which the first season of the show aired, except for “All” for which all interactions are shown.
Violin plot of the average rating per genres for 3 different decades. The outlier dots are interactive and show the rating and the title of the specific TV show.
Each dot shows the season 1 rating for a particular show. Shows are grouped by whether they only had one season (renewed=FALSE) or were renewed for at least a second season (renewed=TRUE). The low p-values for the 2000s and 2010s indicate that shows that were renewed did have better rated first seasons than unrenewed shows, but in the 1990s this was not the case. The statistical power may be worse for the 1990s since there were fewer shows (n=69) than in the 2000s (n=195) or 2010s (n=590).